CIRNet_TIP2022

Runmin Cong, Qinwei Lin, Chen Zhang, Chongyi Li, Xiaochun Cao, Qingming Huang, and Yao Zhao, CIR-Net: Cross-modality interaction and refinement for RGB-D salient object detection, IEEE Transactions on Image Processing, vol. 31, pp. 6800-6815, 2022.

Results of CIR-Net

• Results:

• We provide the resutls of our CIR-Net on six popular RGB-D SOD benchmark datasets, including STEREO797, NLPR, NJUD, DUT, LFSD and SIP.
• The results can be download from: Baidu Cloud (password:1234)

Pytorch Code of CIR-Net:

• Pytorch implementation of CIR-Net
• Pretrained model:
  • We provide our testing code. If you test our model, please download the pretrained model, unzip it, and put the checkpoint CIRNet.pth to CIRNet_cpts/ folder
  • Pretrained model using ResNet50 backbone:Baidu Cloud (password:1234)
  • Pretrained model using VGG16 backbone: Baidu Cloud (password:1234)

Requirements

• Python 3.7
• torch=1.10.1
• torchvision=0.11.2
• opencv-python
• Pillow

Data Preprocessing

• Please download and put the train data to data folder.
• train data can be download from: Baidu Cloud (password:1234)
• test data can be download from: Baidu Cloud (password:1234)

Test

python3 CIRNet_test.py --backbone R50 --test_model CIRNet_R50.pth

Train

python3 CIRNet_train.py --backbone R50

• You can find the results in the test_maps folder

If you use our CIR-Net, please cite our paper:

 @article{crm/tip22/CIRNet,
   title={{CIR-Net}: Cross-modality interaction and refinement for {RGB-D} salient object detection},
   author={Cong, Runmin and Lin, Qinwei and Zhang, Chen and Li, Chongyi and Cao, Xiaochun and Huang, Qingming and Zhao, Yao },
   journal={IEEE Trans. Image Process. },
   volume={31},
   pages={6800-6815},
   year={2022},
  }

Contact Us

If you have any questions, please contact Runmin Cong (rmcong@bjtu.edu.cn) or Qinwei Lin (lqw22@mails.tsinghua.edu.cn).